Phosphoinositide-Regulated Kinases and Phosphoinositide

Jul 20, 2001 - Substrate-Selective Inhibition of Protein Kinase PDK1 by Small Compounds that Bind to the PIF-Pocket Allosteric Docking Site ... Allost...
0 downloads 12 Views 300KB Size
Download PDF
Chem. Rev. 2001, 101, 2365−2380

2365

Phosphoinositide-Regulated Kinases and Phosphoinositide Phosphatases Nick R. Leslie,‡ Ricardo M. Biondi,‡ and Dario R. Alessi*,† MRC Protein Phosphorylation Unit and Division of Signal Transduction Therapy, Department of Life Sciences, University of Dundee, Dundee DD1 5EH, U.K. Received January 19, 2001

Contents I. Introduction II. The Serine/Threonine Kinase PKB Interacts with PtdIns(3,4,5)P3 and PtdIns(3,4)P2 III. PKB Is Activated by Phosphorylation of Thr308 and Ser473 IV. PKB May Be a Key Mediator of Cell Survival, Cell-Cycle Regulation, and Insulin Responsiveness V. Identification of the 3-Phosphoinositide-Dependent Protein Kinase (PDK1) That Activates PKB VI. How Is PKB Phosphorylated at Ser473? VII. Discovery that PDK1 Activates Other Kinases VIII. Mechanism of Regulation of PDK1 Activity IX. Phosphatidylinositol 3-Phosphatase PTEN X. PTEN Protein XI. PTENsA Protein Phosphatase? XII. 5-PhosphatasessSHIP1 and SHIP2 XIII. Concluding Remarks XIV. Acknowledgments XV. References

2365 2365 2367 2367 2369 2370 2371 2373 2374 2374 2376 2376 2377 2377 2377

I. Introduction Stimulation of cells with growth factors, survival factors, and insulin within seconds leads to the recruitment to the plasma membrane of a family of lipid kinases known as Class 1 phosphoinositide 3-kinase (PI 3-kinases, reviewed in refs 1 and 2). In this location PI 3-kinases phosphorylate the glycerophospholipid phosphatidylinositol 4,5-bisphosphate (PtdIns(4,5)P2) at the D-3 position of the inositol ring, converting it to PtdIns(3,4,5)P3, which is then converted to PtdIns(3,4)P2 through the action of the SH2containing inositol phosphatases (SHIP1 and SHIP2) or back to PtdIns(4,5)P2 via the action of the lipid phosphatase termed Phosphatase and TENsin homologue deleted on chromosome TEN (PTEN, Figure 1). Recent evidence indicates that PtdIns(3,4,5)P3 and perhaps PtdIns(3,4)P2 are likely to play a key role in regulating cell growth and survival, differentiation and cytoskeletal changes, as well as mediating many, if not all, of the known physiological responses to * To whom correspondence should be addressed. E-mail: [email protected]. † MRC Protein Phosphorylation Unit. ‡ Division of Signal Transduction Therapy.

insulin. PtdIns(3,4,5)P3 and PtdIns(3,4)P2 mediate cellular effects by interacting with proteins that possess a certain type of pleckstrin homology domain (PH domain).3 In many cases, this recruits these proteins to the plasma membrane. A number of types of PH domain-containing proteins that interact with PtdIns(3,4,5)P3 and/or PtdIns(3,4)P2 have been identified. These include the serine/threonine protein kinases 3-phosphoinositide-dependent protein kinase (PDK1) and protein kinase B (PKB also known as Akt), tyrosine kinases of the Tec family,4,5 numerous adaptor molecules such as the Grb2-associated protein (GAB16), the dual adaptor of phosphotyrosine and 3-phosphoinositides DAPP1,7-10 and the tandem PH domain-containing proteins (TAPP1 and TAPP211) as well as GTP/GDP exchange12,13 and GTPaseactivating proteins14 for the ARF/Rho/Rac family of GTP-binding proteins (Figure 2). In this review we will focus on recent research aimed at understanding the mechanism by which activation of PI 3-kinase, and hence formation of PtdIns(3,4,5)P3, enables PDK1 to phosphorylate and activate a group of serine/threonine protein kinases that belong to the AGC subfamily of protein kinases. These include isoforms of Protein kinase B (also known as Akt),15 p70 ribosomal S6 kinase (S6K),16 serum and glucocorticoid-induced protein kinase (SGK),17-19 and protein kinase C (PKC) isoforms20 (Figure 2). It is believed that the activation of these kinases mediates many of the effects of PI 3-kinase on promoting cell survival and mediating the physiological responses of cells and tissues to insulin. We will also discuss the importance of the PTEN and the SHIP phosphoinositide phosphatases that play key roles in regulating the cellular concentration of PtdIns(3,4,5)P3 and thus regulate the activity of the downstream effector pathways activated by these phosphoinositides.

II. The Serine/Threonine Kinase PKB Interacts with PtdIns(3,4,5)P3 and PtdIns(3,4)P2 There are three isoforms of PKB (termed PKBR, PKBβ, and PKBγ) which possess >85% sequence identity and are widely expressed in human tissues.15 Stimulation of cells with agonists that activate PI 3-kinase invariably induces a large activation of PKB isoforms with a half-time for maximal activation occurring within 1-2 min. There is a strong body of evidence that PKB lies downstream of PI 3-kinase, as the activation of PKB is prevented by inhibitors of PI 3-kinase such as wortmannin or LY294002 or

10.1021/cr000091i CCC: $36.00 © 2001 American Chemical Society Published on Web 07/20/2001

2366 Chemical Reviews, 2001, Vol. 101, No. 8

Nick R. Leslie was born in Chester, England, in 1969. He studied for his First Degree in Genetics at Cambridge University (1991) and his Ph.D. degree at Glasgow University (1995) in David Sherratt’s laboratory, working on site-specific recombination in the bacterial chromosome. His postdoctoral work in the Beatson Institute in Glasgow and the Biochemistry Department, Dundee University, has been in the field of mammalian signal transduction. His current research projects with Peter Downes focus on phosphoinositide signaling and the tumor-suppressor protein PTEN. Nick is a postdoctoral research scientist in the Division of Signal Transduction Therapy, Dundee.

Ricardo M. Biondi was born in Buenos Aires, Argentina, in 1966. He graduated in Biology (1991) and obtained his Ph.D. (1996) degree at the University of Buenos Aires. During a postdoctoral stay at Pasteur Institut in Paris, he developed his interests on how protein kinases really interact with their substrates. For this reason he moved to Dundee and joined the Division of Signal Transduction Therapy as a postdoctoral research scientist, where he has worked with both Dr. Dario R. Alessi and Sir Professor Philip Cohen. In Dundee he has been investigating the involvement of protein kinase−substrate docking interactions as specificity determinants in the insulin/growth factor signaling pathways.

by the overexpression of a dominant negative regulatory subunit of PI 3-kinase.21-23 Furthermore, overexpression of a constitutively active mutant of PI 3-kinase induces PKB activation in unstimulated cells24 (see below) as does deletion of the PTEN phosphatase which results in increased cellular levels of PtdIns(3,4,5)P3.25-29 These observations indicate that PtdIns(3,4,5)P3 and/or PtdIns(3,4)P2 generation in cells is sufficient to induce the activation of PKB isoforms. All PKB isoforms possess an N-terminal PH domain followed by a kinase catalytic domain and then a C-terminal tail. Although pleckstrin homology domains were first noticed in various proteins predicted to have a role in signal transduction in 1993,30-33 it was not until 1995 that it became

Leslie et al.

Dario R. Alessi was born in Strasbourg, France, in 1967. He was educated in Brussels, Belgium, and studied for his First Degree in Biochemistry and Biotechnology at the University of Birmingham (U.K.) and graduated in 1988. He carried out his Ph.D. degree jointly supervised by Professor Ian Trayer at the University of Birmingham and Dr. David Trentham at the National Institute of Medical Research in London. His Ph.D. research focused on the organic synthesis of spin-labeled nucleotide probes to be used in structural investigations of biological problems and was awarded his Ph.D. degree in 1991. In 1991 he moved to the MRC Protein Phosphorylation Unit at the University of Dundee to work under Sir Professor Philip Cohen, studying the role of protein phosphatases and protein kinases regulating diverse physiological processes. In 1997 he became a principal investigator at the MRC Protein Phosphorylation unit in Dundee, and since then the focus of the work in his laboratory has been to study signal transduction processes that are triggered by PI 3-kinase in cells.

Figure 1. Mechanism by which agonists stimulate the formation of PtdIns(3,4,5)P3 and PtdIns(3,4)P2. Agonists bind to their receptors, and this results in the recruitment of PI 3-kinase from the cytosol to the cell membrane, bringing it into the vicinity with its physiological substrate PtdIns(4,5)P2. PtdIns(4,5)P2 is phosphorylated by the p110 catalytic subunit of PI 3-kinase at the D3 position of the inositol ring to generate the second messenger PtdIns(3,4,5)P3. A specific 5-phosphatases termed SHIP1 or SHIP2 converts PtdIns(3,4,5)P3 to PtdIns(3,4)P2, which is also predicted to be a second messenger. A specific 3-phosphatase termed PTEN can convert PtdIns(3,4,5)P3 back to PtdIns(4,5)P2.

apparent that PH domains on certain proteins could function as a phosphoinositide-binding module, enabling proteins to interact with cell membranes (reviewed in ref 3). It was then demonstrated by several groups that PKB or its isolated PH domain bound in vitro with high affinity to PtdIns(3,4,5)P3 and PtdIns(3,4)P2.34,35 Furthermore, it was also shown that following stimulation of cells with agonists that activate PI 3-kinase, PKB was translocated to the plasma membrane, where PtdIns(3,4,5)P3 is located, and that translocation was prevented by inhibitors of PI 3-kinase or by deletion of the PH domain of

Phosphoinositide-Regulated Kinases

Figure 2. Signal transduction pathways known to be triggered by PtdIns(3,4,5)P3 and PtdIns(3,4)P2. Apart from PDK1 and PKB, a number of other proteins possess PH domains that interact with PtdIns(3,4,5)P3 and/or PtdIns(3,4)P2. These include certain adaptor proteins, tyrosine kinases belonging to the TEC family, as well as GDP/GTP exchange proteins and GTPase-activating proteins for the ARF/Rho/Rac family of small GTP binding proteins. A key question for future research is to define the physiological proccess that are regulated by each branch of this pathway.

PKB.36-38 These findings strongly indicate that PKB interacts with PtdIns(3,4,5)P3 and/or PtdIns(3,4)P2 in vivo.

III. PKB Is Activated by Phosphorylation of Thr308 and Ser473 Importantly, PKB is not directly activated by its interaction with either PtdIns(3,4,5)P3 or PtdIns(3,4)P2.35,39 Although some authors claimed that PKB could be activated by binding to PtdIns 3P22 or through its interaction with PtdIns(3,4)P2,34,40,41 these results have not been reproduced and are now not thought to be physiologically relevant.42 As the activated form of PKB was inactivated by serine/ threonine phosphatases,21,43 this indicated that PKB might be activated by phosphorylation on a serine/ threonine residue(s). 32P-cell-labeling experiments revealed that insulin or IGF1 stimulation of cells that potently activates PI 3-kinase induced the phosphorylation of PKBR at two residues, Thr308 and Ser473, and that inhibitors of PI 3-kinase prevented the phosphorylation of both residues.44 Thr308 is located in the “T-loop” (also known as activation loop) between subdomains VII and VIII of the kinase catalytic domain, situated at the same position as the activating phosphorylation sites found in many other protein kinases. Ser473 is located outside of the catalytic domain in a motif Phe-Xaa-Xaa-Phe-Ser-Tyr (where Xaa is any amino acid) that is present in most AGC kinases and is termed the hydrophobic motif. The phosphorylation of PKBR at both Thr308 and Ser473 is likely to be required to activate PKBR maximally as mutation of Thr308 to Ala abolished PKBR activation by insulin or IGF1, whereas mutation of Ser473 to Ala reduced the activation of PKBR by ∼85%. The mutation of both Thr308 or Ser473 to Asp (to mimic the effect of phosphorylation by introducing a negative charge) increased PKBR activity substantially in unstimulated cells, and this mutant could not be further activated by insulin.44 Attachment of a membrane-targetting domain to PKBR resulted in it becoming highly activated in unstimulated cells and induced a maximal phospho-

Chemical Reviews, 2001, Vol. 101, No. 8 2367

rylation of Thr308 and Ser473.36,45 These observations indicated that recruitment of PKB to the membrane of cells could be sufficient for it to become activated at that location by the phosphorylation of Thr308 and Ser473. PKBβ is activated by phosphoryation of the equivalent residues (Thr309 and Ser474), and a splice variant of PKBγ that does not possess a residue equivalent to Ser473 of PKBR (because it terminates at residue 454) is also activated when overexpressed in cells solely via phosphorylation of Thr305 (the residue equivalent to Thr308 of PKBR).46 A longer splice variant of PKBγ has been identified that like the other PKB isoforms possesses a hydrophobic motif, and it is likely that PI 3-kinase activation induces the phosphorylation of the residue equivalent to Ser473 of PKBR in the long splice variant of PKBγ (Ser472).47 There are three reports which indicate that under certain physiological conditions PKB can be activated independently from PI 3-kinase. First, Ca2+ levels in cells have been reported to activate PKB through the Ca2+/calmodulin-dependent protein kinase kinase (CAMKK).48 This study reports that CAMKK phosphorylates PKB on Thr308 in the absence of PtdIns(3,4,5)P3. However, another group found that CAMKK is not capable of inducing the phosphorylation of PKB,49 and others have not been able to induce PKB activation in neuronal, kidney, or fibroblast cell lines by agonists which increase intracellular Ca2+ levels (Shaw, M.; Cohen, P. Unpublished observations). Second, it has also been reported that when PKB is overexpressed in 293 cells, it can be partially activated (∼2-fold) in a PI3K-independent manner by agents that increase cAMP levels.50 Physiological relevance of this requires further investigation, and in this regard we have been unable to detect the activation of endogenously expressed PKB in several cell lines by such stimuli, despite being able to measure large increases in cAMP- and cAMP-dependent protein kinase/PKA activity in these cells (Shaw, M.; Alessi, D. R. Unpublished observations). Third, another study claimed that oxidative stress and heat shock treatment of cells could activate PKB and that this activation was not inhibited by wortmannin.51 However, a subsequent study inicated that activation of PKB by these stressors was inhibited by PI 3-kinase inhibitors.52

IV. PKB May Be a Key Mediator of Cell Survival, Cell-Cycle Regulation, and Insulin Responsiveness Soon after PKB was found to be activated downstream of PI 3-kinase, it was postulated that PKB could function as a mediator of PI 3-kinase-dependent signaling processes in cells. Several groups addressed the question of what roles PKB might play by overexpressing constitutively active mutants of PKB in cell lines. These results demonstrated that overexpression of activated PKB strongly inhibited apoptosis induced by many cellular insults that induce cell death (reviewed in refs 53 and 54). Overexpression of active forms of PKB in insulin-responsive cells induced many processes that are normally activated by insulin such as glycogen synthesis and protein

2368 Chemical Reviews, 2001, Vol. 101, No. 8

Leslie et al.

Table 1a substrate

physiological role of phosphorylation

GSK3R/GSK3β PFK2 PDE3B FKHR BAD mCaspase 9 IKKR eNOS Raf mTOR BRCA1 HTERT IRS P21CIP1/WAF1

stimulates glycogen43 and protein202 synthesis in insulin responsive tissues stimulates glycolysis in heart tissue203 insulin-mediated reduction of cyclicAMP levels in adipocytes73 regulation of transcription and cell survival147,149,204 promotion of cell survival by inhibiting apoptosis69 promotion of cell survival by inhibiting apoptosis164 NF-κB activation by tumor necrosis factor166,167 VEGF or shear-induced activation of eNOs promoting angiogenesis157,158 inhibition of the Raf-MEK-ERK signaling pathway75 stimulation of protein synthesis205,206 regulation nuclear localization165 activation of the human telomerase reverse transcriptase subunit207 inhibition of IRS to activate downstream insulin-mediated pathways208 forces cytoplasmic localization of P21CIP1/WAF1 preventing growth arrest71

a List of proteins that may be phosphorylated physiologically by PKB. The potential role that phosphorylation of these proteins plays in mediating physiological processes is stated. Abreviations: GSK3, glycogen synthase kinase 3; PFK2, 6-phosphofructo2-kinase; PDE 3B, phosphodiesterase 3B; FKHR, forkhead transcription factor; mCaspase9, mouse caspase 9; IKKR, IkB kinaseR; eNOS, endothelial nitric oxide synthase; mTOR, mammalian target of Rapamycin; BRCA1, breast cancer susceptibility gene1; hTERT, human telomerase reverse transcriptase subunit; IRS, insulin receptor subunit, cell-cycle-inhibitor protein. Although there is evidence that PKB phosphorylates mouse caspase 9 at Ser196, as this site is not conserved on the human form of this enzyme,201 the physiological relevance of this observation is not certain at the moment.

synthesis,55,56 stimulated uptake of nutrients such as glucose and amino acids,57,58 and also induced certain changes in gene expression that are induced by insulin and growth factors.57,59,60 Furthermore, the overexpression of active forms of PKB in cells has also been shown to promote differentiation of some cells61-63 while inhibiting the differentiation of others.64,65 It is likely that these processes are mediated by PKB phosphorylating a number of key regulators of cellular function, and much of the work in this area is aimed at identification and characterization of these substrates. It is well established that the minimum sequence motif required in a peptide substrate to be efficiently phosphorylated by PKB lies in the motif Arg-XaaArg-Yaa-Zaa-Ser/Thr-Hyd, where Xaa is any amino acid, Yaa and Zaa are preferably small residues other than glycine, and Hyd is a bulky hydrophobic residue (Phe, Leu).66,67 The requirement for the Arg at positions 3 and 5 residues N-terminal to the phosphorylation site is critical as changing either of these residues even to Lys abolished phosphorylation of peptide substrates by PKB.66 The identification of this PKB phosphorylation motif has led to the identification of over 20 proteins that possess this motif and whose function is known or postulated to be regulated by insulin and other agonists that activate PKB (reviewed in ref 15), and some novel substrates have been proposed by Cantley and colleagues.67 In Table 1 we list the best characterized PKB substrates and state what role that phosphorylation may play in regulating the activity of these proteins, which has also been reviewed elsewhere.15,68-70 A very recent substrate for PKB to be discovered is the inhibitor of the cyclin-dependent kinase inhibitor p21Cip1/WAF1. PKB phosphorylates this protein at Thr145, thereby preventing it from entering the nucleus of cells where it needs to be in order to arrest the growth of cells.71 This mislocalization of a major regulator of cell-cycle progression may be a general mechanism by which various oncogenes such as Her2/neu that activate PKB can enable cells to proliferate under conditions whereby their growth

should be arrested.71,72 Although the presence of an Arg-Xaa-Arg-Xaa-Xaa-Ser/Thr-Hyd motif in a protein is certainly an important determinant in enabling a protein to be phosphorylated by PKB, it is not currently known whether there are other sites on PKB substrates that would act as “docking sites”, enabling PKB to interact with these and hence enhance the rate at which these are phosphorylated. In this regard, it should be noted that several PKB substrates including phosphodiesterase 3B,73 endothelial nitric oxide synthase,74 Raf1,75 and p21Cip1/WAF1 71 have been reported to physically interact with PKB isoforms. It is interesting to note that Thr145 of p21Cip1/WAF1 does not lie in a optimal PKB phosphorylation consensus. Although the Arg residues 5 and 3 residues before the site of phosphorylation are conserved, the residue following Thr145 is a Ser rather than a large hydrophobic residue. One could speculate that the interaction of PKB with p21Cip1/WAF1 (as well as with other PKB substrates), through specific PKB binding motif(s), may be required for them to be phosphorylated efficiently by PKB in vivo. This will require further investigation. It should be noted that although all the PKB substrates listed in Table 1 have been shown to be phosphorylated by PKB in vitro, the evidence that the endogenous forms of these proteins are phosphorylated in vivo in response to agonists that activate PI 3-kinase is in many cases lacking. Furthermore, as no specific PKB inhibitors have been developed, it is very difficult to rule out that the phosphorylation of the PKB substrate protein being studied is mediated by a PKB isoform rather than by another PI 3-kinase-activated protein kinase distinct from PKB (such as SGK). Although dominant negative forms of PKB have been used extensively to dissect the signaling networks that are regulated by PKB, great caution in our opinion should be employed when using these reagents. This is becuase it is not clear what is the mechanism by which dominant negative PKB mutants are functioning in cells. For example, it is possible that dominant negative PKB is working by interacting with and inhibiting the upstream

Phosphoinositide-Regulated Kinases

Chemical Reviews, 2001, Vol. 101, No. 8 2369

protein kinase(s) that activate PKB, and this in turn could also affect the phosphorylation of other AGC kinases, such as SGK. It is also likely that the overexpression of constitutively active or dominant negative mutants of PKB in cell lines for many hours or even days may indirectly evoke physiological responses that are not normally mediated by PKB. For these reasons, the future work in this area will need to be aimed at developing pharmacological reagents and genetic and biochemical approaches that not only identify novel roles for PKB, but also verify which cellular functions are indeed mediated physiologically by PKB.

V. Identification of the 3-Phosphoinositide-Dependent Protein Kinase (PDK1) That Activates PKB An enzyme was purified39,76 and subsequently cloned77,78 that phosphorylated PKBR at Thr308 only in the presence of lipid vesicles containing PtdIns(3,4,5)P3 or PtdIns(3,4)P2 that was termed 3-phosphoinositide-dependent protein kinase-1 (PDK1). It was found to be composed of an N-terminal catalytic domain and a C-terminal PH domain that interacted with PtdIns(3,4,5)P3 and PtdIns(3,4)P2.78,79 The activation of PKB by PDK1 was enhanced over 1000fold in the presence of lipid vesicles containing a low mole fraction of PtdIns(3,4,5)P3 or PtdIns(3,4)P2.77 The activation was stereospecific for the physiological D-enantiomers of these lipids. Furthermore, the naturally occurring form of PtdIns(3,4,5)P3 (the 1-stearoyl, 2-arachidonyl derivative) was significantly more effective than the dipalmitoyl derivative.39,78 It has been postulated that the packing of the unsaturated fatty acids of natural PtdIns(3,4,5)P3 in the phospholipid bilayer of cellular membranes is loose, causing greater exposure of the headgroup of PtdIns(3,4,5)P3 and hence allowing more efficient interactions with PH domain-containing proteins. This may explain why inositol phospholipids, several of which form specific interactions with PH domains or other lipid binding domains, tend to contain fatty acids with unsaturated backbones whereas other nonsignaling phospholipids generally possess more saturated fatty acid moieties. Neither PtdIns(4,5)P2 nor any other inositol phospholipid other than PtdIns(3,4)P2 can replace PtdIns(3,4,5)P3 in the PDK1catalyzed activation of PKB.39,78 Even the presence of a 100-fold molar excess of PtdIns(4,5)P2 over PtdIns(3,4,5)P3 did not prevent PDK1 from phosphorylating and activating PKB (Alessi, D. R. Unpublished results), consistent with the cellular situation where there is typically >100-fold excess of PtdIns(4,5)P2 over PtdIns(3,4,5)P3. The requirement for PtdIns(3,4,5)P3 or PtdIns(3,4)P2 in the PDK1-catalyzed activation of PKB is mediated at least in part by their interaction with the PH domain of PKB. These inositol lipid “second messengers” appear to alter the conformation of PKB, so that Thr308 becomes accessible to PDK1 (Figure 3). These conclusions are supported by the finding that in the absence of PtdIns(3,4,5)P3, full-length PKB is not phosphorylated by PDK1 but removal of the PH domain of PKB results in its phosphorylation and

Figure 3. Mechanism of activation of PKB. Stimulation of cells with agonists that result in the activation of PI 3-kinase trigger the production PtdIns(3,4,5)P3 at the plasma membrane. PKB then interacts with PtdIns(3,4,5)P3 (and/or PtdIns(3,4)P2) through its PH domain and is thus recruited from the cytosol to the plasma membrane. The interaction of PKB with PtdIns(3,4,5)P3 alters its conformation so that Thr308 becomes accessible for phosphorylation by PDK1. PKB is also phosphorylated at the membranes at Ser473 by an as yet uncharacterized protein kinase that is termed here PDK2. Phosphorylation of PKB at Thr308 and Ser473 activates PKB. The interaction of PDK1 with PtdIns(3,4,5)P3 through its PH domain does not activate PDK1 but is thought to play an important role in enabling PDK1 to be located at the membrane so that it can phosphorylate PKB that is recruited to this location. PIP2 stands for PtdIns(4,5)P2, whereas PIP3 indicates PtdIns(3,4,5)P3. (Reprinted with permission from Alessi, D. R. Biochem. Soc. Trans., in press. Copyright 2001 The Biochemical Society.)

activation by PDK1 in vitro.77 Furthermore, a PKBR mutant that cannot interact with PtdIns(3,4,5)P3 is not phosphorylated by PDK1,76 but PtdIns(3,4,5)P3 is still required for the activation of PKBR by a truncated form of PDK1 which lacks the PH domain.77,78 PtdIns(3,4,5)P3 and PtdIns(3,4)P2 interact with the PH domain of PDK1, with somewhat higher affinity than with the PH domain of PKBR.79 It is thought that the interaction of PDK1 with these lipids enables PDK1 and PKB to co-localize at membranes that contain PtdIns(3,4,5)P3 and/or PtdIns(3,4)P2 and that this plays a key role in enabling PDK1 to phosphorylate PKB efficiently. These conclusions are supported by the finding that the rate of activation of PKBR by PDK1 in vitro in the presence of lipid vesicles containing PtdIns(3,4,5)P3 is lowered considerably if the PH domain of PDK1 is deleted. Furthermore, the mutant of PKB that lacks its PH domain is also a very poor substrate for PDK1 compared to wild-type PKB as it is unable to interact with lipid vesicles containing PtdIns(3,4,5)P3. It is controversial as to whether PDK1 translocates to the plasma membrane of cells in response to agonists that activate PI 3-kinase. Two reports79,80 and J. Tavare (University of Bristol, personal communication) indicate that a small proportion of PDK1 is

2370 Chemical Reviews, 2001, Vol. 101, No. 8

associated with the membrane of unstimulated cells, and they do not observe any further translocation of PDK1 to membranes in response to insulin stimulation that activates PI 3-kinase and PKB. However, other groups have reported that PDK1 translocates to cellular membranes in response to agonists that activate PI 3-kinase.81,82 It is possible, as suggested by Van Obberghen and colleagues,82 that unless the basal levels of PtdIns(3,4,5)P3 in unstimulated cells are maintained at a very low level, it may be hard to observe translocation of PDK1 to membrane of cells as it possesses a high affinity for PtdIns(3,4,5)P3, and thus, a significant amount of PDK1 could co-localize with low amounts of PtdIns(3,4,5)P3 at the plasma membranes of unstimulated cells. Indeed, there is evidence that PDK1 is likely to be located at cell membranes of unstimulated cells as the expression of a membrane-targeted PKB construct in such cells is active and fully phosphorylated at Thr308.36,45 This observation indicates that a significant amount of endogenous PDK1 as well as PtdIns(3,4,5)P3 must be present at the membrane of these cells to enable PKB that is fixed at this location to become phosphorylated and activated.

VI. How Is PKB Phosphorylated at Ser473? A major outstanding question is to characterize the mechanism by which PKB is phosphorylated at its hydrophobic motif. There is currently considerable controversy on whether PKB is phosphorylated at this residue by a distinct upstream kinase that has not yet been identified (sometimes referred to as PDK2) or whether PKB, after it becomes phosphorylated at Thr308, is able to phosphorylate itself at Ser473. In this section we will review the evidence for and against each of these models. We39,46,77,83 and others76,78 originally postulated that PKB would not be phosphorylated at Ser473 by autophosphorylation or by PDK1 as it was found that PDK1 only phosphorylated PKB significantly at Thr308 in vitro and that PKB phosphorylated at Thr308 (which is partially activated) was not able to autophosphorylate itself at Ser473 even in the presence of lipid vesicles containing PtdIns(3,4,5)P3. This was also supported by the finding discussed above that a catalytically inactive mutant of PKB or a mutant of PKB in which Thr308 is changed to either Ala or Glu was still phosphorylated on Ser473 following stimulation of 293 cells with IGF1.44 However, some caution is required when interpreting this latter result as it could be argued that either the kinase dead PKB transfected into cells is not 100% “dead” and still could possess the ability to autophosphorylate itself at Ser473 or that this phosphorylation could be mediated by the wild-type endogenous PKB present in these cells. This notion was recently challenged by Toker and Newton, who reported that PDK1 in vitro was capable of inducing the PKB to phosphorylate itself at Ser473.84 This analysis was carried out using a very sensitive phosphospecific antibody that recognizes PKB phosphorylated at Ser473, and thus, it is uncertain whether the phosphorylation observed at Ser473 in these experiments occurred at trace levels

Leslie et al.

or at a significant stoichiometry. Further evidence that PKB may be able to phosphorylate itself at Ser473, at least in vitro, is the finding that PKBR when incubated with PtdIns(3,4,5)P3 and PDK1 complexed to a C-terminal fragment of PKC-related kinase-2, termed PDK1 interacting fragment (PIF), became phosphorylated at both Thr308 and Ser473.83 Although we originally interpreted this result to imply that PDK1 was phosphorylating both Thr308 and Ser473 of PKB, the subsequent observation that catalytically inactive mutants of PKB are only phosphorylated at Thr308 and not at Ser473 by PDK1 when complexed to PIF85 may indicate that PKB is phosphorylating itself at Ser473 under these conditions. It is therefore possible that the PIF peptide might be able to bind to PKB and following phosphorylation of Thr308 enable PKB to phosphorylate Ser473 by a mechanism that has not been characterized. It should be noted that this conclusion is not definitive as it could be argued that the catalytically inactive PKB could be misfolded and therefore not recognized by PDK1 complexed to PIF. Furthermore, we have also not been able to detect a significant interaction between PKB and PIF by co-immunoprecipitation or surface plasmon resonance binding studies, although it could be argued that these proteins are able to interact with each other weakly. Mouse embryonic stem (ES) cells in which both copies of the PDK1 gene were disrupted to prevent the expression of PDK1 have recently been generated.86 These cells (termed PDK1-/- ES cells) were viable and proliferated at a similar rate to the wildtype ES cells expressing PDK1 (termed PDK1+/+ ES cells). IGF1-stimulation of PDK1-/- ES cells activated PI 3-kinase to a greater level than in the control PDK1+/+ ES cells, but PKB was not activated significantly in the PDK1-/- ES cells whereas it was activated robustly in the control PDK1+/+ ES cells. As expected, IGF1-stimulation of PDK1-/- ES cells failed to induce phosphorylation of PKB at Thr308 in its T-loop, providing the first genetic evidence in mammalian cells that PDK1 mediates this phosphorylation. Importantly, however, PKB was still phosphorylated at Ser473 in response to IGF1 stimulation, and this was inhibited by the PI 3-kinase inhibitors wortmannin and LY294002.86 This indicated that, at least in PDK1-/- ES cells, PDK1 is not required for the phosphorylation of PKB at Ser473. One explanation for these data is that a kinase distinct from PDK1 and PKB can mediate phosphorylation of Ser473. An alternative possibility could be that PKB might be regulated by a PIF-like molecule in vivo that, in the absence of PDK1 and phosphorylation of PKB at Thr308, can bind to PKB and enable phosphorylation of Ser473. It is possible that phosphorylation of PKB at Ser473 could be regulated differently in other cell types or in response to different agonists. It will therefore be important in the future to prepare other cell lines which lack PDK1 and to establish the effect that this has on Ser473 phosphorylation in response to a variety of agonists. The availability of a specific PDK1 inhibitor (which has not yet been developed) would also be a extremely valuable tool with which to corroborate

Phosphoinositide-Regulated Kinases

data regarding the inability of PDK1 to regulate the phosphorylation of PKB at its hydrophobic motif in PDK1 knockout cell lines. In this regard, it should be noted that Hemmings and colleagues found that the nonspecific kinase inhibitor staurosporine (which was presumed, but not demonstrated, to inhibit PDK1) prevented the phosphorylation of PKBR at Thr308 but not Ser473 in 293 cells.87 This provides pharmacological evidence that PDK1 is not essential for the phosphorylation of PKBR at Ser473 in these cells. Blenis and colleagues made the very interesting observation that addition of N-R-tosyl-L-phenylalanylchloromethyl ketone (TPCK) to cells prevented insulin-stimulated PKB activation and its phosphorylation at both Thr308 and Ser473, without affecting activation of PI 3-kinase.88 This study indicates that TPCK may inhibit PDK1 (and phosphorylation of PKBR at Ser473) in vivo, but as TPCK does not inhibit PDK1 in vitro from phosphorylating PKB,88 further investigations are required to establish how this drug is inhibiting PKB activation in cells. Interestingly, other than PDK1 or PKB, three other protein kinases have been reported to mediate the phosphorylation of PKB at its hydrophobic motif, namely, the integrin-linked kinase (ILK), MAPKAP kinase-2, and conventional PKC isoforms. Dedhar and colleagues89 reported that ILK was capable of directly phosphorylating PKB at Ser473 as well as being able to interact with PtdIns(3,4,5)P3 despite not possessing a PH domain. We (Alessi, D. R.; Dowler, S. Unpublished data) and others90 have not been able to reproduce these results, and more recently it has been reported that ILK may regulate the phosphorylation of PKB at Ser473 by an indirect mechanism as the overexpression of both a wild-type and a catalytically inactive mutant of ILK was capable of inducing phosphorylation of PKB at Ser473 in unstimulated cells.90 MAPKAP kinase-2 is activated by phosphorylation by the p38 MAP kinase following exposure of cells to cellular stresses and cytokines. We originally purified MAPKAP kinase-2 from rabbit skeletal muscle as an activity that could phosphorylate PKBR at Ser473 in vitro.44 However, as an inhibitor of p38 termed SB203580 that prevented MAPKAPK2 activation in cells did not affect phosphorylation of PKBR at Ser473 in many cell lines52 including in ES cells (Lizcano, J. Unpublished results) and agonists such as sodium arsenite and UV irradiation that activated MAPKAP kinase-2 did not induce phosphorylation of PKBR52 at Ser473, this was taken as evidence that MAPKAPK2 did not phosphorylate this residue of PKBR physiologically. However, a recent study demonstrates that in neutrophils stimulated with FMLP, SB 203580 is capable of inhibiting phosphorylation of PKB at Ser473 at low concentrations of the SB inhibitor (